Our faculty member, Assoc. Prof. Zeynep Tuna Değer, along with her colleagues, has published a new article in the Journal of Building Engineering, introducing a rapid assessment method of the overall damage level of post-earthquake buildings based on component images and deep learning.
Frequent seismic events significantly impact building structures, reducing their safety and serviceability. This study proposes a deep learning (DL)-based approach using convolutional neural networks (CNNs) to classify structural damage from component images.
The proposed method first trains two classification models for identifying component types and damage levels. After an earthquake, images of affected buildings are collected and classified. By combining the weights of various component types and damage levels, the method allows for the calculation of overall structural condition scores and grades. This provides a comprehensive evaluation of structural damage after seismic events and extends to building portfolios, offering valuable guidance for personnel allocation, resource distribution, rescue operations, and subsequent maintenance measures.
To read the full article: https://doi.org/10.1016/j.jobe.2024.111380